Governance and Scalability: Circles of trust and federated platforms

Written by: Cristina Ampatzidou & Irina Shapiro

How large a circle (of trust) can be?

Independent publishing platforms are often small in size and human resources and usually not able to afford advertising and other forms of paid visibility to increase their audiences. Amateur Cities, Open! and Open Set seem to share much of the same concerns: we are all small, independent platforms, following a slow publishing process, focused on high quality, yet addressing a limited audience. As such, joining forces and bringing our communities together is seen as a way to amplify our reach, while at the same time providing additional interesting content to our audiences. In the framework of the Making Public project, we are prototyping a tool that creates relations between content items located in different publishing platforms, and provides recommendations to readers. Based on a set of tags, Clickpong suggests a selection of articles from the network’s platforms, that are displayed as recommendations on the front-end of other participating platforms.

Clickpong creates a closed, federated, trust-based peer-to-peer network, where content value is legitimized by the other platforms participating in the network. The network is closed because it consists of a limited number of participants and entry of new members is approved by the existing ones; federated because each participating platform maintains an absolute autonomy of its editorial strategy; trust-based because each platform concedes to feature links to other platforms; and peer-to-peer because it is a distributed network, where all participating platforms have an equal position and overview of the network. This of course assumes that the editors of each platform are familiar with the values, publishing strategies and content of the other platforms and are confident that they relate to their own and will be meaningful also to their audiences.

Because trust is such a fundamental part of establishing this network, the architecture of the tool also facilitates transparency, so that the editors can know based on which criteria a recommendation was created and can vote it as more or less successful leading to the gradual improvement of the algorithm. However, the transparency of the tool and familiarization of the participating platforms are not all the key elements that establish trust within the network.

Meaning and Levels of Trust

The transparency of the publishing strategies and algorithms of Clickpong build up positive expectations shared by the parties involved in the network, and are mainly based on: the reputation, values and assurance in the integrity of the involved platforms; understanding and appreciation of the platforms’ audiences; as well as appreciation of the content, the engaging with which will be a meaningful time investment for all the platforms. The step to make from the positive expectations to trust in the network’s collaboration lies not in the increasing familiarities or transparencies, but rather in a decision of taking a vulnerable position of opening up each other’s content for cross sharing. This decision means to take a risk of allowing our content to be decontextualized on the front-end of other platforms, while appearing in combination with articles that might alternate or, even further, create a critical reading of the featured article. In other words, the decision of trusting the platforms in a situation that might lead to the regretful or problematic outcomes for the platform that opens its content for cross-sharing.

What is it to regret? A very sensitive (or risky) moment in the discussion of our group was the control over the selection and appearance of the articles on each other’s platforms. At some point there was an idea discussed  to create articles recommendations that appear directly on the front end of the peer-platforms without editors’ pre-selection, however, with an option to be deleted later on. That step required a lot of confidence in each other’s content quality and compatibility of it, which is possible to achieve in a very small scale of circle of collaborators and by a very close evaluation of the shared articles. But is it possible to have this assurance by using an automated tool for a larger networks of partners? Or this is where the risky decision of trusting each other becomes essential for the network? Even within our limited experiment with a small number of publishers, we already encountered a relevant issue, when we discussed the possibility of automated recommendations, without the manual approval of a platform editor. UN Studio, which was also part of the initial team, heavily objected to automated recommendations. As a publishing platform, they focus on promoting research and technologies related to the projects developed by the associated architectural practice. For them it was too risky to accept external content without moderation, while having an obligation to commercial partners. Eventually UNStudio stepped out of the project as its focus and needs diverged from the rest of the group. The idea of direct appearance of the articles has not gone through to the final prototype of Clickpong, however, there is risk in how an article from one platform is used, combined with / on other platforms.

In this context, trust is seen as the basis for sharing the content within the network, as a risk taking decision in collaboration, as well as a decision of solidarity with the ethical position and attitude of the organizations and individuals who are involved in the network. Through this understanding of trust, it’s interesting to look at the quality that comes with the cross referencing each other content.

Federation and Peer Governance

One of the most interesting aspects of Clickpong is the possibility to scale up and include more platforms. Being a federated network means that each platform maintains a total autonomy with regards to their editorial strategy and the content they decide to publish, however joining the Clickpong network also means that each platform concedes to host content of the other participating platforms. Obviously, as with many peer-to-peer networks, scaling up comes together with certain questions concerning the governance of the structure. What is the ideal number of platforms that can be part of it? How is it decided which new platforms can join in? Does that influence how links are established? UN Studio’s decision to leave the project prompted a lot of discussion among the group on the governance structure of the tool and the trust issues that might come at stake with scaling up. Finding common ground among three publishers can be easy, but what happens when more platforms join in? Can a consensus-based governance structure be maintained? How easy it is to keep track and get to know other publishers’ content in order to decide whether they “fit the profile” of the Clickpong?

Peer governance is a bottom-up mode of participative decision-making that is characterized by the fact that all participants in a peer network have an equal ability to contribute, an heterarchical organization, and overview of the whole network. With only three remaining publishers on board, we have been able to manage the tool development by meeting physically, discussing and arriving at consensual decisions. In case of a larger numbers of participants, possibly geographically dispersed, a more formalized organisational structure would be required to maintain the good functionality of the network. In that case the question becomes: How to maintain balance between hierarchy and decentralization? The existence of some sort of centralized leadership in many peer-to-peer projects, often referred to as a ‘benevolent dictator’, is rather common, so it is not unreasonable to think that the initiators and early adopters of Clickpong would take up this role in case of a scale up. One of the core tasks of managing such a project is indeed to maintain the quality of the product or function, and prevent its degradation from the addition of low quality contributions. Indeed, for all publishers currently participating in Clickpong, ensuring the high quality of content, both their own and recommended one is of major importance. In that sense, even though in principle every platform is potentially equally able to participate and contribute to the network, the people who manage the project should be equally free to reject platforms for joining Clickpong.